Re: Extended Prefetching using Asynchronous IO - proposal and patch - Mailing list pgsql-hackers

Thanks Fujii ,   that is a bug   --   an #ifdef  USE_PREFETCH is missing in heapam.c
   (maybe several)
I will fix it in the next patch version.

I also appreciate it is not easy to review the patch.
There are really 4 (or maybe 5) parts :
  
     .   async io (librt functions)
     .   buffer management   (allocating, locking and pinning etc)
     .   scanners making prefetch calls
     .   statistics

    and the autoconf input program

I will see what I can do.   Maybe putting an indicator against each modified file?

I am currently working on two things :
  .   alternative way for non-originator of an aio_read to wait on completion
             (LWlock instead of polling the aiocb)
      This was talked about in several earlier posts and Claudio is also working on something there
  .   package up my benchmark

Cheers    John

----------------------------------------
> Date: Fri, 20 Jun 2014 04:21:19 +0900
> Subject: Re: [HACKERS] Extended Prefetching using Asynchronous IO - proposal and patch
> From: masao.fujii@gmail.com
> To: johnlumby@hotmail.com
> CC: pgsql-hackers@postgresql.org; klaussfreire@gmail.com
>
> On Mon, Jun 9, 2014 at 11:12 AM, johnlumby <johnlumby@hotmail.com> wrote:
>> updated version of patch compatible with git head of 140608,
>> (adjusted proc oid and a couple of minor fixes)
>
> Compilation of patched version on MacOS failed. The error messages were
>
> gcc -O0 -Wall -Wmissing-prototypes -Wpointer-arith
> -Wdeclaration-after-statement -Wendif-labels
> -Wmissing-format-attribute -Wformat-security -fno-strict-aliasing
> -fwrapv -g -I../../../../src/include -c -o heapam.o heapam.c
> heapam.c: In function 'heap_unread_add':
> heapam.c:362: error: 'struct HeapScanDescData' has no member named
> 'rs_Unread_Pfetched_next'
> heapam.c:363: error: 'struct HeapScanDescData' has no member named
> 'rs_Unread_Pfetched_count'
> heapam.c:369: error: 'struct HeapScanDescData' has no member named
> 'rs_Unread_Pfetched_count'
> heapam.c:375: error: 'struct HeapScanDescData' has no member named
> 'rs_Unread_Pfetched_base'
> heapam.c:381: error: 'struct HeapScanDescData' has no member named
> 'rs_Unread_Pfetched_base'
> heapam.c:387: error: 'struct HeapScanDescData' has no member named
> 'rs_Unread_Pfetched_next'
> heapam.c:405: error: 'struct HeapScanDescData' has no member named
> 'rs_Unread_Pfetched_count'
> heapam.c: In function 'heap_unread_subtract':
> heapam.c:419: error: 'struct HeapScanDescData' has no member named
> 'rs_Unread_Pfetched_next'
> heapam.c:425: error: 'struct HeapScanDescData' has no member named
> 'rs_Unread_Pfetched_count'
> heapam.c:434: error: 'struct HeapScanDescData' has no member named
> 'rs_Unread_Pfetched_base'
> heapam.c:442: error: 'struct HeapScanDescData' has no member named
> 'rs_Unread_Pfetched_base'
> heapam.c:452: error: 'struct HeapScanDescData' has no member named
> 'rs_Unread_Pfetched_next'
> heapam.c:453: error: 'struct HeapScanDescData' has no member named
> 'rs_Unread_Pfetched_next'
> heapam.c:454: error: 'struct HeapScanDescData' has no member named
> 'rs_Unread_Pfetched_next'
> heapam.c:456: error: 'struct HeapScanDescData' has no member named
> 'rs_Unread_Pfetched_count'
> heapam.c: In function 'heapgettup':
> heapam.c:944: error: 'struct HeapScanDescData' has no member named
> 'rs_pfchblock'
> heapam.c:716: warning: unused variable 'ix'
> heapam.c: In function 'heapgettup_pagemode':
> heapam.c:1243: error: 'struct HeapScanDescData' has no member named
> 'rs_pfchblock'
> heapam.c:1029: warning: unused variable 'ix'
> heapam.c: In function 'heap_endscan':
> heapam.c:1808: error: 'struct HeapScanDescData' has no member named
> 'rs_Unread_Pfetched_base'
> heapam.c:1809: error: 'struct HeapScanDescData' has no member named
> 'rs_Unread_Pfetched_base'
> make[4]: *** [heapam.o] Error 1
> make[3]: *** [heap-recursive] Error 2
> make[2]: *** [access-recursive] Error 2
> make[1]: *** [install-backend-recurse] Error 2
> make: *** [install-src-recurse] Error 2
>
>
> Huge patch is basically not easy to review. What about simplifying the patch
> by excluding non-core parts like the change of pg_stat_statements, so that
> reviewers can easily read the patch? We can add such non-core parts later.
>
> Regards,
>
> --
> Fujii Masao